Short answer
A 2026 study from Writer and Workplace Intelligence found that 29% of employees admitted to sabotaging their company’s AI strategy, with the figure rising to 44% among Gen Z workers. As ITPro reported, sabotage included refusing to use AI tools, bypassing training, using unapproved tools, and mishandling sensitive data. The lesson for leaders is awkward but useful: AI adoption does not fail only because the tool is bad. It can fail because people do not trust the rollout.
What happened?
In April 2026, research from Writer and Workplace Intelligence found a striking gap between executive AI enthusiasm and employee behavior. The survey, which covered workers across the U.S., U.K., and Europe, found that 29% of employees admitted to undermining their company’s AI strategy in some form. Among Gen Z employees, the figure reportedly rose to 44%.
Fast Company reported that the study included 2,400 workers, split between C-suite executives and employees ranging from individual contributors to managers. The forms of sabotage varied. Some employees refused to use AI tools. Some bypassed training. Some used unapproved tools. Some entered sensitive company information into public AI systems.
ITPro’s coverage framed the behavior less as cartoonish rebellion and more as a response to poorly managed AI integration. Employees cited concerns including job displacement, diminished creativity, increased workload, and lack of trust in how AI was being introduced.
That framing matters. “Sabotage” is a dramatic word, and it risks making employees sound like tiny office villains with spreadsheet capes. The more useful interpretation is that many workers are anxious, confused, unconvinced, or quietly routing around the official process.
That is still a risk. It is just a more human one.
Why should leaders care?
AI transformation is now a boardroom priority, but adoption happens at desk level. Employees decide whether to use the tool, trust the output, disclose AI involvement, follow data rules, complete training, or find their own workaround.
That makes AI adoption a culture challenge as much as a technology program.
If employees fear replacement, they may resist. If they think AI tools are being forced on them without explanation, they may disengage. If the approved tool is poor but a public tool works better, they may use the public tool quietly. If training feels like compliance theater, they may skip it. If leadership says “AI will empower everyone” while cutting headcount, employees may hear the second message louder.
This matters for cyber and human risk because poor adoption creates unsafe behavior. People may paste sensitive data into unapproved tools. They may rely on AI outputs without checking. They may hide AI use from managers. They may bypass controls to prove a point or get work done. They may refuse to use approved tools, leaving teams with fragmented practices and no clear view of risk.
A failed AI rollout is not only a productivity problem. It can become a data, security, quality, and trust problem.
The human risk behind AI resistance
Employees do not resist technology in a vacuum. They respond to incentives, pressure, uncertainty, and culture.
If AI is introduced as a cost-cutting weapon, people will protect themselves. If policies are unclear, people will improvise. If managers do not understand the tools, teams will set their own norms. If employees are punished for asking difficult questions, they will stop asking in public and start experimenting in private.
That is where shadow AI and AI sabotage overlap. One person refuses to use the approved tool. Another uses a banned tool because it works better. Another submits AI-generated work without disclosure. Another enters sensitive data into a public system because the company told everyone to “use AI more” but never explained what safe use actually means.
These behaviors are not always malicious. Many are signs that the organization has not earned trust or created a safe path. Human risk management helps leaders see that gap before it becomes an incident.
The uncomfortable truth is that employees are part of the AI control environment. Treat them like obstacles and the control environment gets weird fast.
What organizations should do now
Organizations should start by treating AI adoption as a change-management and culture program, not just a software rollout. Employees need to understand why AI is being introduced, what it will and will not be used for, how it affects their roles, and what safe behavior looks like.
Be clear about data rules. Which tools are approved? What data can be used? What must never be entered into public AI systems? When should AI-generated output be reviewed? When does AI use need to be disclosed? These rules should be practical, role-specific, and easy to find.
Managers need support too. They are the people employees will watch most closely. If managers are confused, dismissive, or secretly using tools differently from the policy, teams will notice. Train managers to discuss AI openly, handle concerns, and reinforce safe use without making people feel foolish.
Organizations should also involve employees in the rollout. Let people test tools, give feedback, identify risky workflows, and shape use cases. People are more likely to support AI when they have a voice in how it changes their work.
Finally, measure the human side. Do employees trust the AI strategy? Do they understand the rules? Are they hiding AI use? Are they afraid to disclose mistakes? Are teams using unapproved tools because the official path is too slow? These are human risk indicators, and they deserve the same seriousness as technical metrics.
The Cybermaniacs take
AI rollout sabotage is a human risk management story because it shows how technology adoption can fail through behavior, trust, and culture.
Cyber culture matters when employees decide whether to follow AI rules, report unsafe use, challenge outputs, protect sensitive data, or admit they are using tools unofficially. Leaders cannot govern what they cannot see, and they cannot see what employees are afraid to say.
For Cybermaniacs, this is exactly why AI governance needs to be tied to human risk management. Organizations need role-based learning, culture measurement, safe-use guidance, leadership communication, and practical nudges that help people use AI safely in real work. A policy alone will not close the gap between executive ambition and employee behavior.
AI automation may be powerful, but culture still gets the deciding vote. Rude of it, perhaps. Historically consistent.
FAQ
What is AI rollout sabotage?
AI rollout sabotage refers to employee behaviors that undermine an organization’s AI strategy. Reported examples include refusing to use approved tools, bypassing AI training, using unapproved AI tools, or mishandling sensitive company data.
Why would employees sabotage AI adoption?
Employees may fear job displacement, distrust leadership, dislike poorly implemented tools, feel excluded from the rollout, worry about creativity or workload, or find official AI policies unclear or unrealistic.
How does this create security risk?
If employees hide AI use, use unapproved tools, paste sensitive data into public platforms, or rely on AI outputs without review, organizations can lose visibility and control over data, quality, and compliance.
How can companies reduce AI resistance?
Communicate clearly, involve employees early, provide approved tools that work well, train managers, create practical data rules, address job-impact concerns honestly, and build a culture where people can raise concerns safely.
Why is this human risk management?
Because AI adoption depends on behavior. Human risk management helps organizations understand trust, pressure, hidden workarounds, data-handling habits, and the culture needed for safe AI use.